RUNOFF AND EROSION FOLLOWING A PRESCRIBED FIRE ON A SAGEBRUSH-STEPPE RANGELAND IN IDAHO, USA

. A study was carried out to compare runoff and erosion from natural rainfall on plots that had been treated with prescribed fire to unburned plots on a sagebrush-steppe rangeland in Southeast Idaho, U.S. Prescribed fire on rangeland sites is intended to maintain healthy shrub-steppe ecosystems but sometimes results in undesirable consequences such as increased runoff and soil erosion. Information is lacking on plot-scale erosion studies from natural precipitation in this ecosystem. Such plot-scale studies are needed to better understand sediment sources (uplands or channels) to support the management and modeling of rangeland watersheds. In this study, there were two treatments: four prescribed burn plots and three adjacent unburned control plots. Runoff and erosion were measured from natural rainfall for water years (WY) 2004 to 2010 following a prescribed burn in October 2003. Runoff and erosion were also measured from nearby unburned plots for WY 2005-2010. Ground cover on the burned plots averaged 57% (standard error (s.e. = 4%) in the summer of 2004, compared to 95% (s.e. = 5%) on the unburned plots. By the end of the study in 2010, ground cover had increased to 81% (s.e. = 4%) on the burned plots but decreased to 74% (s.e. = 5%) on the unburned plots. Annual runoff averaged 6 mm (s.e. = 7) from four burned plots, compared to 34 mm (s.e. = 8.7) from three unburned plots. In WY 2006, high rates of runoff from snowmelt on the unburned plots resulted in 122 mm (s.e. = 16) of runoff compared to only 12 mm (s.e. = 14) of runoff from the burned plots. An analysis of variance showed significant differences in runoff due to either precipitation (p = 0002) or year (p = 0.004) and treatment (burned vs. unburned; p=0.03). There were also significant differences in seasonal runoff (p = 0.05), as 90% of the measured runoff occurred in the spring, with all large runoff events associated with snowmelt. Erosion on the burned plots averaged 233 kg ha-1 (s.e. = 21) compared to 133 kg ha-1 (s.e. = 29) on the unburned plots. From two years after the burn and for the remainder of the study, there were no significant differences in erosion between burned and unburned plots (p < 0.05). Future studies are needed to link upland runoff and erosion with channel deposition, erosion, and sediment delivery, and more detailed studies on erosion associated with snowmelt on rangelands are needed to aid in the development of watershed modeling tools. Future studies should include observations of plant community regeneration in addition to ground cover, runoff, and sediment delivery.